Inferring social structure from digital trace data : applications to organizations and markets
Abstract/Contents
- Abstract
- This thesis explores the use of large, unstructured, digital-trace data toward inferring aspects of latent social structure in one market context and one organizational context. It is structured as three separate papers (chapters), where, although starkly di erent in setting, topic, and methods, each exhibits this overarching theme in some way. In Chapter 1, the context is cultural product markets, where quality assessments are largely subjective; consumers are subject to rapid satiation upon repeated exposure to the same stimuli; and the forms and attributes of the products o ered are determined principally by cultural norms. Although organizational sociologists often test theories in these contexts as empirical settings, as is argued in that chapter, those theories are often impoverished by the coarseness of the data they typically work with. In Chapter 1's analysis, digital-trace data (in the form of crowd-sourced identifications of product attributes) is applied toward identifying implicit category spanning and characterizing the products to a richer extent than is typical in the literature. This ultimately allows more nuanced hypotheses of strategic positioning to be tested. In Chapter 2, a natural language processing model is developed for quantifying the extent to which an input text is "elaborate" vs "compact", i.e., for quantifying the text's linguistic code. The notion derives from a concept in sociolinguistics which asserts that the linguistic codes used between two speakers reveals the nature of the relationship between them, with regard to the subject matter at hand. Elaborated code is accessible to outsiders, and its use implies an assumption of a lack of shared contextual background, whereas compact code implies an assumption of insider status. The model combines multiple methods in an ensemble approach, and post training is validated using two distinct assessments. In Chapter 3 this model is applied on digital trace data of communication in an organizational setting, toward an investigation of gender-based returns to communication styles. The analysis revisits an empirical inconsistency in the literature, leveraging the trace data to explore consequences of speech patterns at scale. I nd signals of distinct patterns of communication between men and women, which correlate with disparate outcomes in performance evaluations and hiring.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2023; ©2023 |
| Publication date | 2023; 2023 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Oshotse, Abraham |
|---|---|
| Degree supervisor | Carroll, Glenn |
| Degree supervisor | Goldberg, Amir |
| Thesis advisor | Carroll, Glenn |
| Thesis advisor | Goldberg, Amir |
| Thesis advisor | Sorensen, Jesper B, 1967- |
| Degree committee member | Sorensen, Jesper B, 1967- |
| Associated with | Stanford University, Graduate School of Business |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | Abraham O. Oshotse. |
|---|---|
| Note | Submitted to the Graduate School of Business. |
| Thesis | Thesis Ph.D. Stanford University 2023. |
| Location | https://purl.stanford.edu/gn346mg0472 |
Access conditions
- Copyright
- © 2023 by Abraham Oshotse
- License
- This work is licensed under a Creative Commons Attribution 3.0 Unported license (CC BY).
Also listed in
Loading usage metrics...